How to configure the `upstream-timeout` plugin

How do I configure the upstream-timeout plugin?

The upstream-timeout plugin overrides the timeouts configured at a service level.

The configuration parameters are:

config.read_timeout: Reading from upstream timeout threshold in milliseconds

config.send_timeout: Send to upstream timeout threshold in milliseconds

config.connect_timeout: Connect to upstream timeout threshold in milliseconds

The plugin can be enabled globally, on a service or a route, same as other plugins. For example, to enable the plugin on a route with a read_timeout set to override any service-configured read_timeout:

curl -X POST http://<kong-admin-api>:8001/routes/{route_id}/plugins --header "kong-admin-token: <token>" \
--data "name=upstream-timeout" \
--data "config.read_timeout=4000"

If the target Route was created without an explicit protocols field, Kong Gateway defaults it to ["https"] only (not ["http","https"]), so a plain http:// request to that route returns 426 Please use HTTPS protocol regardless of this plugin. Set protocols explicitly (e.g. ["http","https"]) when creating the Route if you need plain-HTTP access.

To enable the plugin on Kubernetes

apiVersion: configuration.konghq.com/v1
kind: KongClusterPlugin
metadata:
  name: global-upstream
  annotations:
    kubernetes.io/ingress.class: kong
  labels:
    global: "true"
config:
  read_timeout: 4000
plugin: upstream-timeout